Cancer incidence other than gastric cancer in pernicious anemia: a systematic review with meta-analysis

Background: Pernicious anemia (PA) has been associated with an increased risk of cancer development. While an increased incidence of gastric cancer and carcinoids was demonstrated, evidence is still quite conflicting on the risk for other cancer types. The aim of this study was to systematically review the available literature on PA and the development of gastrointestinal-other than gastric cancers (GI-other than GCs) and non-gastrointestinal cancers (non-GICs) to estimate their incidence-rates. Methods: According to PRISMA, we identified studies on PA patients reporting the incidence of GI-other than GCs and non-GICs. Studies quality was evaluated using the Newcastle-Ottawa Quality Assessment Scale. Meta-analysis on annual cancer incidence rates was performed. Results: Twenty studies met the eligibility criteria. On a total PA population of 82,257 patients, cumulative cancer incidence rates per 100,000 person-years (PY) were 13.0, 8.2, and 17.5 for all cancers, GI-other than GCs and non-GICs, respectively, with calculated pooled cancer incidence-rates (PY) of 0.27 and 0.23, respectively. Compared to published data on the general population, meta-analysis showed an overall cancer relative risk in PA as 0.68 (95% CI: 0.48-0.95). PA patients showed a lower cancer relative risk for colorectum (0.14: 0.098-0.19), breast (0.17: 0.13-0.22), liver (0.18: 0.13-0.24), esophagus (0.26: 0.18-0,37), lung (0.26: 0.21-0.31), thyroid (0.34: 0,21-0,57), ovary (0.45: 0.30-0.67), non-melanoma skin cancer (0.56: 0.50-0.61), kidney (0.61: 0.43-0.86); in contrast, PA patients showed a higher relative risk for biliary tract cancer (1.81: 1.21 to 2.70) and multiple myeloma (2.83: 1.76-4.55), Hodgkin lymphoma (3.0: 1.35-6.68), non Hodgkin lymphoma (2.08: 1.58-2.75), and leukemia (1.56: 1.16-2.12). Conclusions: This systematic review showed an overall lower cancer relative risk in PA patients, but nearly a twofold relative risk of biliary tract cancers and a twofold-three fold relative risk of hematological malignancies. Further high quality studies are needed to confirm the higher risk of these specific cancers in PA patients.